Re: primary keys and dictionary overhead
Date: Mon, 24 Oct 2011 11:22:20 +0530
Unique key and PK are somewhat explained in this OTN thread.
in nutshell the difference in unique and PK is
Unique = No duplicate , but allow any number of NULLs
Primary key = No duplicate and no NULLs
I think, this sort of design need arises when the application has planty of UPDATES than INSERTS..
i.e. it will come with some default unique data. which will be updated in future.eg CRM data, PO schedules etc..
On 23 October 2011 09:49, Sidney Chen <huanshengchen_at_gmail.com> wrote:
> two question:
> 1. what's your motivation to add unique constraints into primary keys? it's
> from business request, or it's because you believe it the constraint will
> benifit the performance?
> 2. There must be a reason why previous guy decide to pull out the
> constraints, I think there should be a benchmark report to address this
> decision, see if you can get the document. James Mole did not mention
> constraint related design in his exceptional book "Scaling Oracle 8i", but
> he did emphasize the importance of the benchmark in a whole chapter.
> > --
> > http://www.freelists.org/webpage/oracle-l
> Sidney Chen
-- ============================================= TRUTH WINS AT LAST, DO NOT FORGET TO SMILE TODAY ============================================= -- http://www.freelists.org/webpage/oracle-lReceived on Mon Oct 24 2011 - 00:52:20 CDT